CIOInsider India Magazine

When storytelling meets cutting-edge automation, the outcome is Newton, a physics engine that simulates robotic movements in real-world settings. Created through a three way partnership between Disney Research, NVIDIA, and Google DeepMind, Newton represents the power that can transform the future of robotics, not only in the entertainment industry but in many more sectors as well.

Debuted at NVIDIA’s AI Conference, the key product of this partnership is Newton—an advanced, open-source physics engine designed to replicate realistic, high-fidelity interactions for robots. Through NVIDIA’s robust Warp framework, Newton is believed to enable robots to learn, adapt, and move with a sense of realism and emotion.

But there is another star that grabbed the limelight of the show and that is an adorable, compact robot named Blue. Inspired by Star Wars droids, Blue impressed the audience by showing a bunch of expressiveness and agility. Not only that, the robot engaged in an interaction with NVIDIA CEO Jensen Huang by communicating solely through beeps and gestures. Blue’s interaction appeared to be more of a character than mere programming product.

What truly sets this apart is the synergy of NVIDIA’s simulation and AI capabilities, Google DeepMind’s expertise in deep learning, and Disney’s exceptional storytelling and robotics experience. Together, trio has developed a product that empowers robots to not only execute tasks but also to forge connections with people.

The Technical Framework Within Newton
Newton’s central system comprises the NVIDIA Warp framework, a CUDA-based acceleration library that facilitates the writing of high-performance, GPU-accelerated code directly in Python. Combining parallel processing capabilities of NVIDIA GPUs, Warp allows roboticists to train models at unmatched speeds. This is what helps Newton to conduct real-time simulations, help robots fine-tune their movements more accurately, adapt to environments more efficiently, and excel in both precise industrial tasks and interactive entertainment applications.

Unmatched Speed
Regarding simulation speed, Newton takes a large piece of the pie here. Created by Google DeepMind, MuJoCo-Warp merges the physics engine of MuJoCo with the acceleration features of Warp, leading to simulation speed enhancements of over 70 times for robotic manipulation tasks, with some specific tasks experiencing up to a 100 times increase in speed. This progress implies that complex physical interactions, such as a robotic hand using a tool, can be simulated with remarkable efficiency, greatly shortening the development cycle and reducing the computational expenses of robotic systems.

What Jobs Does Disney Plan to Rely on Newton for
For long it’s known that Disney has been engulfed in introducing Star Wars-themed robots to its global parks. It has even performed numerous controlled experiments of these droids. Now that Newton is here, Disney is most likely to hire the robot at its various theme park locations starting next year from Orlando to Tokyo to Paris—bringing realistic robotic characters to life in ways that were once limited to science fiction.
Newton could help enhance the physical capabilities, expressiveness, and intelligence of other animatronics like Donald Duck and Scrooge McDuck at its theme parks.

Furthermore, Disney Research intends to use Newton as the power source for its upcoming entertainment robots, which include the BDX droids inspired by Star Wars. These droids, meant to engage with visitors in Disney theme parks, were displayed during the GTC 2025 keynote, showcasing realistic movements and actions.
Disney Research could be among the first users of Newton, incorporating it into their robotic character platform to create expressive and interactive robots for theme parks and entertainment. To conclude GTC 2025, NVIDIA CEO Jensen Huang presented Disney’s BDX droids, Star Wars-inspired robots that utilize Newton’s physics simulations for realistic movements and gestures.

Through NVIDIA’s Isaac robotics platform, DeepMind’s AI reinforcement learning, and Disney’s renowned storytelling skills, we are perhaps in for a revolution of interactive robotic characters that seem more lifelike than ever.

In addition to entertainment, Newton’s features also apply to industrial robotics, AI-powered humanoid assistants, and advanced manufacturing systems, all of which necessitate high-fidelity simulations prior to real-world implementation.

Newton Employs Differential Physics
One of the primary obstacles in robotics development is the sim-to-real gap, which refers to the disparity between robot behavior in simulations and their actions in the physical world. Newton addresses this issue by employing differentiable physics, enabling robots to learn from physics-driven simulations and modify their movements to suit real-world conditions.

Additionally, Newton is built to be adaptable, allowing developers to incorporate custom solvers, connect multiphysics environments, and simulate interactions with flexible objects such as food, fabric, or sand. This versatility makes Newton an effective tool for training robots in a variety of scenarios.

Meet Newton's Other Friend - Isaac GR00T N1
In addition to Newton, NVIDIA launched Isaac GR00T N1, an open-source foundation model for humanoid robots that enables them to learn and behave like humans. This model empowers robots to comprehend context, engage in reasoning, and adapt to their surroundings—a significant advancement in the field of robotics.

While Newton powers Disney's duck-type robots, GR00T N1 may be utilized for more sophisticated humanoid animatronics, potentially leading to the development of lifelike Disney character robots in theme parks.

NVIDIA aims to launch an initial, open-source version of Newton in 2025, encouraging the global robotics community to take part in its development. This initiative is anticipated to boost progress in robotics, providing researchers and developers around the world with access to advanced simulation tools.

Also Read: Is Starlink Bringing an Internet Revolution to India?

As Newton continues to develop, it is set to become a fundamental component in the progression of humanoid robotics. The initial version, expected to launch later this year, will mark the beginning of a period where AI-driven, fast robotic simulations are commonplace. The partnership between NVIDIA, Google DeepMind, and Disney Research not only brings forth innovative tools and methodologies to robotics but also sets the stage for the broader use of robots in everyday life. Newton's open-source format can encourage creativity among researchers and developers around the world, propelling the robotics field to new levels.